Data modeling and NLP-based scoring method to assess the relevance of environmental regulatory announcements

Abstract

The constantly growing body of global environmental legislation necessitates that corporate environmental compliance managers frequently assess the relevance of new regulations and regulation revisions for each of their sites. Companies are pressured to streamline and automate this crucial task through digital workflows and specialized IT-based assistance systems. This has recently piqued the interest of researchers working in different disciplines, such as intelligent systems, machine learning, and natural language processing. The article describes the latest results of our long-term research program on IT-based support for corporate compliance management, offering insights for these, and other disciplines. The context and the main aspects of environmental regulation announcements and the relevance assessment task are analyzed. An extensive conceptual data model is developed that serves as a foundation for tailoring a generic method to perform a relevance assessment that considers site-specific individual environmental compliance facts. The method uses heuristic data operations and various text processing techniques from the field of natural language understanding. In order to exemplify the method, two application scenarios are described in which the relevance of new waste management directives are assessed for a multi-site production company.